StringUtils.wrapText will not split surrogate pair. (#103)

mikesamuel · web-flow · commit 5338313a54c6 · 2020-11-22T19:48:05.000+01:00
Previously,

    StringUtils.wrapText(".\uD800\uDC00.", 2)
    //                           ^ index 2 points here

would insert a `&lt;br/&gt;` between two paired surrogates.

Now, it shifts the insertion point to a whole code-point boundary.
diff --git a/java-diff-utils/src/main/java/com/github/difflib/text/StringUtils.java b/java-diff-utils/src/main/java/com/github/difflib/text/StringUtils.java
@@ -21,10 +21,10 @@
 final class StringUtils {
 
     /**
-     * Replaces all opening an closing tags with <code>&lt;</code> or <code>&gt;</code>.
+     * Replaces all opening and closing tags with <code>&lt;</code> or <code>&gt;</code>.
      *
      * @param str
-     * @return
+     * @return str with some HTML meta characters escaped.
      */
     public static String htmlEntites(String str) {
         return str.replace("<", "&lt;").replace(">", "&gt;");
@@ -61,7 +61,17 @@ public static String wrapText(String line, int columnWidth) {
         StringBuilder b = new StringBuilder(line);
 
         for (int count = 0; length > widthIndex; count++) {
-            b.insert(widthIndex + delimiter * count, "<br/>");
+            int breakPoint = widthIndex + delimiter * count;
+            if (Character.isHighSurrogate(b.charAt(breakPoint - 1)) &&
+                Character.isLowSurrogate(b.charAt(breakPoint))) {
+              // Shift a breakpoint that would split a supplemental code-point.
+              breakPoint += 1;
+              if (breakPoint == b.length()) {
+                // Break before instead of after if this is the last code-point.
+                breakPoint -= 2;
+              }
+            }
+            b.insert(breakPoint, "<br/>");
             widthIndex += columnWidth;
         }
 
diff --git a/java-diff-utils/src/test/java/com/github/difflib/text/StringUtilsTest.java b/java-diff-utils/src/test/java/com/github/difflib/text/StringUtilsTest.java
@@ -49,6 +49,8 @@ public void testWrapText_String_int() {
         assertEquals("te<br/>st", StringUtils.wrapText("test", 2));
         assertEquals("tes<br/>t", StringUtils.wrapText("test", 3));
         assertEquals("test", StringUtils.wrapText("test", 10));
+        assertEquals(".\uD800\uDC01<br/>.", StringUtils.wrapText(".\uD800\uDC01.", 2));
+        assertEquals("..<br/>\uD800\uDC01", StringUtils.wrapText("..\uD800\uDC01", 3));
     }
 
     @Test

Original file line number	Diff line number	Diff line change
`@@ -49,6 +49,8 @@ public void testWrapText_String_int() {`
`49`	`49`	`assertEquals("te<br/>st", StringUtils.wrapText("test", 2));`
`50`	`50`	`assertEquals("tes<br/>t", StringUtils.wrapText("test", 3));`
`51`	`51`	`assertEquals("test", StringUtils.wrapText("test", 10));`
	`52`	`+ assertEquals(".\uD800\uDC01<br/>.", StringUtils.wrapText(".\uD800\uDC01.", 2));`
	`53`	`+ assertEquals("..<br/>\uD800\uDC01", StringUtils.wrapText("..\uD800\uDC01", 3));`
`52`	`54`	`}`
`53`	`55`
`54`	`56`	`@Test`